1 ;;; org-feed.el --- Add RSS feed items to Org files
3 ;; Copyright (C) 2009 Free Software Foundation, Inc.
5 ;; Author: Carsten Dominik <carsten at orgmode dot org>
6 ;; Keywords: outlines, hypermedia, calendar, wp
7 ;; Homepage: http://orgmode.org
10 ;; This file is part of GNU Emacs.
12 ;; GNU Emacs is free software: you can redistribute it and/or modify
13 ;; it under the terms of the GNU General Public License as published by
14 ;; the Free Software Foundation, either version 3 of the License, or
15 ;; (at your option) any later version.
17 ;; GNU Emacs is distributed in the hope that it will be useful,
18 ;; but WITHOUT ANY WARRANTY; without even the implied warranty of
19 ;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
20 ;; GNU General Public License for more details.
22 ;; You should have received a copy of the GNU General Public License
23 ;; along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>.
24 ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
28 ;; This module allows to create and change entries in an Org-mode
29 ;; file triggered by items in an RSS feed. The basic functionality is
30 ;; geared toward simply adding new items found in a feed as outline nodes
31 ;; to an Org file. Using hooks, arbitrary actions can be triggered for
32 ;; new or changed items.
34 ;; Selecting feeds and target locations
35 ;; ------------------------------------
37 ;; This module is configured through a single variable, `org-feed-alist'.
38 ;; Here is an example, using a notes/tasks feed from reQall.com.
40 ;; (setq org-feed-alist
42 ;; "http://www.reqall.com/user/feeds/rss/a1b2c3....."
43 ;; "~/org/feeds.org" "ReQall Entries")
45 ;; With this setup, the command `M-x org-feed-update-all' will
46 ;; collect new entries in the feed at the given URL and create
47 ;; entries as subheadings under the "ReQall Entries" heading in the
48 ;; file "~/org-feeds.org". Each feed should normally have its own
49 ;; heading - however see the `:drawer' parameter.
51 ;; Besides these standard elements that need to be specified for each
52 ;; feed, keyword-value pairs can set additional options. For example,
53 ;; to de-select transitional entries with a title containing
55 ;; "reQall is typing what you said",
57 ;; you could use the `:filter' argument:
59 ;; (setq org-feed-alist
61 ;; "http://www.reqall.com/user/feeds/rss/a1b2c3....."
62 ;; "~/org/feeds.org" "ReQall Entries"
63 ;; :filter my-reqall-filter)))
65 ;; (defun my-reqall-filter (e)
66 ;; (if (string-match "reQall is typing what you said"
67 ;; (plist-get e :title))
71 ;; See the docstring for `org-feed-alist' for more details.
74 ;; Keeping track of previously added entries
75 ;; -----------------------------------------
77 ;; Since Org allows you to delete, archive, or move outline nodes,
78 ;; org-feed.el needs to keep track of which feed items have been handled
79 ;; before, so that they will not be handled again. For this, org-feed.el
80 ;; stores information in a special drawer, FEEDSTATUS, under the heading
81 ;; that received the input of the feed. You should add FEEDSTATUS
82 ;; to your list of drawers in the files that receive feed input:
84 ;; #+DRAWERS: PROPERTIES LOGBOOK FEEDSTATUS
89 ;; org-feed.el is based on ideas by Brad Bozarth who implemented a
90 ;; similar mechanism using shell and awk scripts.
97 (declare-function url-retrieve-synchronously
"url" (url))
98 (declare-function xml-node-children
"xml" (node))
99 (declare-function xml-get-children
"xml" (node child-name
))
100 (declare-function xml-get-attribute
"xml" (node attribute
))
101 (declare-function xml-get-attribute-or-nil
"xml" (node attribute
))
103 (defgroup org-feed nil
104 "Options concerning RSS feeds as inputs for Org files."
108 (defcustom org-feed-alist nil
109 "Alist specifying RSS feeds that should create inputs for Org.
110 Each entry in this list specified an RSS feed tat should be queried
111 to create inbox items in Org. Each entry is a list with the following items:
113 name a custom name for this feed
115 file the target Org file where entries should be listed
116 headline the headline under which entries should be listed
118 Additional arguments can be given using keyword-value pairs. Many of these
119 specify functions that receive one or a list of \"entries\" as their single
120 argument. An entry is a property list that describes a feed item. The
121 property list has properties for each field in the item, for example `:title'
122 for the `<title>' field and `:pubDate' for the publication date. In addition,
123 it contains the following properties:
125 `:item-full-text' the full text in the <item> tag
126 `:guid-permalink' t when the guid property is a permalink
128 Here are the keyword-value pair allows in `org-feed-alist'.
131 The name of the drawer for storing feed information. The default is
132 \"FEEDSTATUS\". Using different drawers for different feeds allows
133 several feeds to target the same inbox heading.
135 :filter filter-function
136 A function to select interesting entries in the feed. It gets a single
137 entry as parameter. It should return the entry if it is relevant, or
140 :template template-string
141 The default action on new items in the feed is to add them as children
142 under the headline for the feed. The template describes how the entry
143 should be formatted. If not given, it defaults to
144 `org-feed-default-template'.
146 :formatter formatter-function
147 Instead of relying on a template, you may specify a function to format
148 the outline node to be inserted as a child. This function gets passed
149 a property list describing a single feed item, and it should return a
150 string that is a properly formatted Org outline node of level 1.
152 :new-handler function
153 If adding new items as children to the outline is not what you want
154 to do with new items, define a handler function that is called with
155 a list of all new items in the feed, each one represented as a property
156 list. The handler should do what needs to be done, and org-feed will
157 mark all items given to this handler as \"handled\", i.e. they will not
158 be passed to this handler again in future readings of the feed.
159 When the handler is called, point will be at the feed headline.
161 :changed-handler function
162 This function gets passed a list of all entries that have been
163 handled before, but are now still in the feed and have *changed*
164 since last handled (as evidenced by a different sha1 hash).
165 When the handler is called, point will be at the feed headline.
168 This function gets passed a buffer, and should return a list of entries,
169 each being a property list containing the `:guid' and `:item-full-text'
170 keys. The default is `org-feed-parse-rss-feed'; `org-feed-parse-atom-feed'
173 :parse-entry function
174 This function gets passed an entry as returned by the parse-feed
175 function, and should return the entry with interesting properties added.
176 The default is `org-feed-parse-rss-entry'; `org-feed-parse-atom-entry'
180 (list :value
("" "http://" "" "")
182 (string :tag
"Feed URL")
183 (file :tag
"File for inbox")
184 (string :tag
"Headline for inbox")
187 (list :inline t
:tag
"Filter"
189 (symbol :tag
"Filter Function"))
190 (list :inline t
:tag
"Template"
192 (string :tag
"Template"))
193 (list :inline t
:tag
"Formatter"
195 (symbol :tag
"Formatter Function"))
196 (list :inline t
:tag
"New items handler"
198 (symbol :tag
"Handler Function"))
199 (list :inline t
:tag
"Changed items"
200 (const :changed-handler
)
201 (symbol :tag
"Handler Function"))
202 (list :inline t
:tag
"Parse Feed"
204 (symbol :tag
"Parse Feed Function"))
205 (list :inline t
:tag
"Parse Entry"
207 (symbol :tag
"Parse Entry Function"))
210 (defcustom org-feed-drawer
"FEEDSTATUS"
211 "The name of the drawer for feed status information.
212 Each feed may also specify its own drawer name using the `:drawer'
213 parameter in `org-feed-alist'.
214 Note that in order to make these drawers behave like drawers, they must
215 be added to the variable `org-drawers' or configured with a #+DRAWERS
218 :type
'(string :tag
"Drawer Name"))
220 (defcustom org-feed-default-template
"\n* %h\n %U\n %description\n %a\n"
221 "Template for the Org node created from RSS feed items.
222 This is just the default, each feed can specify its own.
223 Any fields from the feed item can be interpolated into the template with
224 %name, for example %title, %description, %pubDate etc. In addition, the
225 following special escapes are valid as well:
227 %h the title, or the first line of the description
228 %t the date as a stamp, either from <pubDate> (if present), or
231 %u,%U like %t,%T, but inactive time stamps
232 %a A link, from <guid> if that is a permalink, else from <link>"
234 :type
'(string :tag
"Template"))
236 (defcustom org-feed-save-after-adding t
237 "Non-nil means save buffer after adding new feed items."
241 (defcustom org-feed-retrieve-method
'url-retrieve-synchronously
242 "The method to be used to retrieve a feed URL.
243 This can be `curl' or `wget' to call these external programs, or it can be
244 an Emacs Lisp function that will return a buffer containing the content
245 of the file pointed to by the URL."
248 (const :tag
"Internally with url.el" url-retrieve-synchronously
)
249 (const :tag
"Externally with curl" curl
)
250 (const :tag
"Externally with wget" wget
)
251 (function :tag
"Function")))
253 (defcustom org-feed-before-adding-hook nil
254 "Hook that is run before adding new feed items to a file.
255 You might want to commit the file in its current state to version control,
260 (defcustom org-feed-after-adding-hook nil
261 "Hook that is run after new items have been added to a file.
262 Depending on `org-feed-save-after-adding', the buffer will already
267 (defvar org-feed-buffer
"*Org feed*"
268 "The buffer used to retrieve a feed.")
271 (defun org-feed-update-all ()
272 "Get inbox items from all feeds in `org-feed-alist'."
274 (let ((nfeeds (length org-feed-alist
))
275 (nnew (apply '+ (mapcar 'org-feed-update org-feed-alist
))))
276 (message "%s from %d %s"
277 (cond ((= nnew
0) "No new entries")
278 ((= nnew
1) "1 new entry")
279 (t (format "%d new entries" nnew
)))
281 (if (= nfeeds
1) "feed" "feeds"))))
284 (defun org-feed-update (feed &optional retrieve-only
)
285 "Get inbox items from FEED.
286 FEED can be a string with an association in `org-feed-alist', or
287 it can be a list structured like an entry in `org-feed-alist'."
288 (interactive (list (org-completing-read "Feed name: " org-feed-alist
)))
289 (if (stringp feed
) (setq feed
(assoc feed org-feed-alist
)))
291 (error "No such feed in `org-feed-alist"))
293 (let ((name (car feed
))
296 (headline (nth 3 feed
))
297 (filter (nth 1 (memq :filter feed
)))
298 (formatter (nth 1 (memq :formatter feed
)))
299 (new-handler (nth 1 (memq :new-handler feed
)))
300 (changed-handler (nth 1 (memq :changed-handler feed
)))
301 (template (or (nth 1 (memq :template feed
))
302 org-feed-default-template
))
303 (drawer (or (nth 1 (memq :drawer feed
))
305 (parse-feed (or (nth 1 (memq :parse-feed feed
))
306 'org-feed-parse-rss-feed
))
307 (parse-entry (or (nth 1 (memq :parse-entry feed
))
308 'org-feed-parse-rss-entry
))
309 feed-buffer inbox-pos new-formatted
310 entries old-status status new changed guid-alist e guid olds
)
311 (setq feed-buffer
(org-feed-get-feed url
))
312 (unless (and feed-buffer
(bufferp (get-buffer feed-buffer
)))
313 (error "Cannot get feed %s" name
))
315 (throw 'exit feed-buffer
))
316 (setq entries
(funcall parse-feed feed-buffer
))
317 (ignore-errors (kill-buffer feed-buffer
))
319 (save-window-excursion
320 (setq inbox-pos
(org-feed-goto-inbox-internal file headline
))
321 (setq old-status
(org-feed-read-previous-status inbox-pos drawer
))
322 ;; Add the "handled" status to the appropriate entries
323 (setq entries
(mapcar (lambda (e)
324 (setq e
(plist-put e
:handled
329 ;; Find out which entries are new and which are changed
331 (if (not (plist-get e
:handled
))
333 (setq olds
(nth 2 (assoc (plist-get e
:guid
) old-status
)))
336 (plist-get e
:item-full-text
))
340 ;; Parse the relevant entries fully
341 (setq new
(mapcar parse-entry new
)
342 changed
(mapcar parse-entry changed
))
346 (setq new
(delq nil
(mapcar filter new
))
347 changed
(delq nil
(mapcar filter new
))))
349 (when (not (or new changed
))
350 (message "No new items in feed %s" name
)
353 ;; Get alist based on guid, to look up entries
356 (mapcar (lambda (e) (list (plist-get e
:guid
) e
)) new
)
357 (mapcar (lambda (e) (list (plist-get e
:guid
) e
)) changed
)))
359 ;; Construct the new status
363 (setq guid
(plist-get e
:guid
))
365 ;; things count as handled if we handle them now,
366 ;; or if they were handled previously
367 (if (assoc guid guid-alist
) t
(plist-get e
:handled
))
368 ;; A hash, to detect changes
369 (sha1 (plist-get e
:item-full-text
))))
372 ;; Handle new items in the feed
376 (goto-char inbox-pos
)
377 (funcall new-handler new
))
378 ;; No custom handler, do the default adding
379 ;; Format the new entries into an alist with GUIDs in the car
382 (lambda (e) (org-feed-format-entry e template formatter
))
385 ;; Insert the new items
386 (org-feed-add-items inbox-pos new-formatted
))
388 ;; Handle changed items in the feed
389 (when (and changed-handler changed
)
390 (goto-char inbox-pos
)
391 (funcall changed-handler changed
))
393 ;; Write the new status
394 ;; We do this only now, in case something goes wrong above, so
395 ;; that would would end up with a status that does not reflect
396 ;; which items truely have been handled
397 (org-feed-write-status inbox-pos drawer status
)
399 ;; Normalize the visibility of the inbox tree
400 (goto-char inbox-pos
)
403 (org-cycle-hide-drawers 'children
)
405 ;; Hooks and messages
406 (when org-feed-save-after-adding
(save-buffer))
407 (message "Added %d new item%s from feed %s to file %s, heading %s"
408 (length new
) (if (> (length new
) 1) "s" "")
410 (file-name-nondirectory file
) headline
)
411 (run-hooks 'org-feed-after-adding-hook
)
415 (defun org-feed-goto-inbox (feed)
416 "Go to the inbox that captures the feed named FEED."
418 (list (if (= (length org-feed-alist
) 1)
420 (org-completing-read "Feed name: " org-feed-alist
))))
421 (if (stringp feed
) (setq feed
(assoc feed org-feed-alist
)))
423 (error "No such feed in `org-feed-alist"))
424 (org-feed-goto-inbox-internal (nth 2 feed
) (nth 3 feed
)))
427 (defun org-feed-show-raw-feed (feed)
428 "Show the raw feed buffer of a feed."
430 (list (if (= (length org-feed-alist
) 1)
432 (org-completing-read "Feed name: " org-feed-alist
))))
433 (if (stringp feed
) (setq feed
(assoc feed org-feed-alist
)))
435 (error "No such feed in `org-feed-alist"))
437 (org-feed-update feed
'retrieve-only
))
438 (goto-char (point-min)))
440 (defun org-feed-goto-inbox-internal (file heading
)
441 "Find or create HEADING in FILE.
442 Switch to that buffer, and return the position of that headline."
445 (goto-char (point-min))
446 (if (re-search-forward
447 (concat "^\\*+[ \t]+" heading
"[ \t]*\\(:.*?:[ \t]*\\)?$")
449 (goto-char (match-beginning 0))
450 (goto-char (point-max))
451 (insert "\n\n* " heading
"\n\n")
452 (org-back-to-heading t
))
455 (defun org-feed-read-previous-status (pos drawer
)
456 "Get the alist of old GUIDs from the entry at POS.
457 This will find DRAWER and extract the alist."
460 (let ((end (save-excursion (org-end-of-subtree t t
))))
461 (if (re-search-forward
462 (concat "^[ \t]*:" drawer
":[ \t]*\n\\([^\000]*?\\)\n[ \t]*:END:")
464 (read (match-string 1))
467 (defun org-feed-write-status (pos drawer status
)
468 "Write the feed STATUS to DRAWER in entry at POS."
471 (let ((end (save-excursion (org-end-of-subtree t t
)))
473 (if (re-search-forward (concat "^[ \t]*:" drawer
":[ \t]*\n")
476 (goto-char (match-end 0))
477 (delete-region (point)
479 (and (re-search-forward "^[ \t]*:END:" nil t
)
480 (match-beginning 0)))))
481 (outline-next-heading)
482 (insert " :" drawer
":\n :END:\n")
483 (beginning-of-line 0))
484 (insert (pp-to-string status
)))))
486 (defun org-feed-add-items (pos entries
)
487 "Add the formatted items to the headline as POS."
491 (unless (looking-at org-complex-heading-regexp
)
492 (error "Wrong position"))
493 (setq level
(org-get-valid-level (length (match-string 1)) 1))
494 (org-end-of-subtree t t
)
495 (skip-chars-backward " \t\n")
496 (beginning-of-line 2)
498 (while (setq entry
(pop entries
))
499 (org-paste-subtree level entry
'yank
))
500 (org-mark-ring-push pos
))))
502 (defun org-feed-format-entry (entry template formatter
)
503 "Format ENTRY so that it can be inserted into an Org file.
504 ENTRY is a property list. This function adds a `:formatted-for-org' property
505 and returns the full property list.
506 If that property is already present, nothing changes."
508 (funcall formatter entry
)
509 (let (dlines fmt tmp indent time name
510 v-h v-t v-T v-u v-U v-a
)
511 (setq dlines
(org-split-string (or (plist-get entry
:description
) "???")
513 v-h
(or (plist-get entry
:title
) (car dlines
) "???")
514 time
(or (if (plist-get entry
:pubDate
)
515 (org-read-date t t
(plist-get entry
:pubDate
)))
517 v-t
(format-time-string (org-time-stamp-format nil nil
) time
)
518 v-T
(format-time-string (org-time-stamp-format t nil
) time
)
519 v-u
(format-time-string (org-time-stamp-format nil t
) time
)
520 v-U
(format-time-string (org-time-stamp-format t t
) time
)
521 v-a
(if (setq tmp
(or (and (plist-get entry
:guid-permalink
)
522 (plist-get entry
:guid
))
523 (plist-get entry
:link
)))
524 (concat "[[" tmp
"]]\n")
528 (goto-char (point-min))
529 (while (re-search-forward "%\\([a-zA-Z]+\\)" nil t
)
530 (setq name
(match-string 1))
532 ((member name
'("h" "t" "T" "u" "U" "a"))
533 (replace-match (symbol-value (intern (concat "v-" name
))) t t
))
534 ((setq tmp
(plist-get entry
(intern (concat ":" name
))))
537 (beginning-of-line 1)
538 (when (looking-at (concat "^\\([ \t]*\\)%" name
"[ \t]*$"))
539 (setq tmp
(org-feed-make-indented-block
540 tmp
(org-get-indentation))))))
541 (replace-match tmp t t
))))
544 (defun org-feed-make-indented-block (s n
)
545 "Add indentation of N spaces to a multiline string S."
546 (if (not (string-match "\n" s
))
549 (org-split-string s
"\n")
550 (concat "\n" (make-string n ?\
)))))
552 (defun org-feed-skip-http-headers (buffer)
553 "Remove HTTP headers from BUFFER, and return it.
554 Assumes headers are indeed present!"
555 (with-current-buffer buffer
557 (goto-char (point-min))
558 (search-forward "\n\n")
559 (delete-region (point-min) (point))
562 (defun org-feed-get-feed (url)
563 "Get the RSS feed file at URL and return the buffer."
565 ((eq org-feed-retrieve-method
'url-retrieve-synchronously
)
566 (org-feed-skip-http-headers (url-retrieve-synchronously url
)))
567 ((eq org-feed-retrieve-method
'curl
)
568 (ignore-errors (kill-buffer org-feed-buffer
))
569 (call-process "curl" nil org-feed-buffer nil
"--silent" url
)
571 ((eq org-feed-retrieve-method
'wget
)
572 (ignore-errors (kill-buffer org-feed-buffer
))
573 (call-process "wget" nil org-feed-buffer nil
"-q" "-O" "-" url
)
575 ((functionp org-feed-retrieve-method
)
576 (funcall org-feed-retrieve-method url
))))
578 (defun org-feed-parse-rss-feed (buffer)
579 "Parse BUFFER for RSS feed entries.
580 Returns a list of entries, with each entry a property list,
581 containing the properties `:guid' and `:item-full-text'."
582 (let (entries beg end item guid entry
)
583 (with-current-buffer buffer
585 (goto-char (point-min))
586 (while (re-search-forward "<item>" nil t
)
588 end
(and (re-search-forward "</item>" nil t
)
589 (match-beginning 0)))
590 (setq item
(buffer-substring beg end
)
591 guid
(if (string-match "<guid\\>.*?>\\(.*?\\)</guid>" item
)
592 (org-match-string-no-properties 1 item
)))
593 (setq entry
(list :guid guid
:item-full-text item
))
597 (nreverse entries
))))
599 (defun org-feed-parse-rss-entry (entry)
600 "Parse the `:item-full-text' field for xml tags and create new properties."
602 (insert (plist-get entry
:item-full-text
))
603 (goto-char (point-min))
604 (while (re-search-forward "<\\([a-zA-Z]+\\>\\).*?>\\([^\000]*?\\)</\\1>"
606 (setq entry
(plist-put entry
607 (intern (concat ":" (match-string 1)))
609 (goto-char (point-min))
610 (unless (re-search-forward "isPermaLink[ \t]*=[ \t]*\"false\"" nil t
)
611 (setq entry
(plist-put entry
:guid-permalink t
))))
614 (defun org-feed-parse-atom-feed (buffer)
615 "Parse BUFFER for Atom feed entries.
616 Returns a list of entries, with each entry a property list,
617 containing the properties `:guid' and `:item-full-text'.
619 The `:item-full-text' property actually contains the sexp
620 formatted as a string, not the original XML data."
621 (with-current-buffer buffer
623 (let ((feed (car (xml-parse-region (point-min) (point-max)))))
627 :guid
(car (xml-node-children (car (xml-get-children entry
'id
))))
628 :item-full-text
(prin1-to-string entry
)))
629 (xml-get-children feed
'entry
)))))
631 (defun org-feed-parse-atom-entry (entry)
632 "Parse the `:item-full-text' as a sexp and create new properties."
633 (let ((xml (car (read-from-string (plist-get entry
:item-full-text
)))))
634 ;; Get first <link href='foo'/>.
635 (setq entry
(plist-put entry
:link
637 (car (xml-get-children xml
'link
))
639 ;; Add <title/> as :title.
640 (setq entry
(plist-put entry
:title
641 (car (xml-node-children
642 (car (xml-get-children xml
'title
))))))
643 (let* ((content (car (xml-get-children xml
'content
)))
644 (type (xml-get-attribute-or-nil content
'type
)))
647 ((string= type
"text")
648 ;; We like plain text.
649 (setq entry
(plist-put entry
:description
(car (xml-node-children content
)))))
650 ((string= type
"html")
651 ;; TODO: convert HTML to Org markup.
652 (setq entry
(plist-put entry
:description
(car (xml-node-children content
)))))
653 ((string= type
"xhtml")
654 ;; TODO: convert XHTML to Org markup.
655 (setq entry
(plist-put entry
:description
(prin1-to-string (xml-node-children content
)))))
657 (setq entry
(plist-put entry
:description
(format "Unknown '%s' content." type
)))))))
662 ;; arch-tag: 0929b557-9bc4-47f4-9633-30a12dbb5ae2
663 ;;; org-feed.el ends here